Insight into the antibacterial drug design and architectural mechanism of peptide recognition from the E. faecium peptide deformylase structure


  • Ki Hyun Nam,

    1. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-713, Korea
    Search for more papers by this author
    • Ki Hyun Nam and Jung Il Ham contributed equally to this work

  • Jung Il Ham,

    1. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-713, Korea
    2. Department of Medicine Laboratory, National Cancer Center, Gyeounggi 410-769, Korea
    Search for more papers by this author
    • Ki Hyun Nam and Jung Il Ham contributed equally to this work

  • Amit Priyadarshi,

    1. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-713, Korea
    2. Biomedical Research Center, Korea Institute of Science and Technology, Seoul 136-791, Korea
    Search for more papers by this author
  • Eunice Eunkyeong Kim,

    1. Biomedical Research Center, Korea Institute of Science and Technology, Seoul 136-791, Korea
    Search for more papers by this author
  • NamHyun Chung,

    1. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-713, Korea
    Search for more papers by this author
  • Kwang Yeon Hwang

    Corresponding author
    1. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-713, Korea
    • Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-701, Korea
    Search for more papers by this author
    • Ki Hyun Nam and Jung Il Ham contributed equally to this work


Following translation initiation, the formyl group of the growing polypeptide is removed by peptide deformylase (PDF) to yield the mature protein.1, 2 PDF is essential for bacterial growth, thus making it an attractive target for the design of new antibiotic drugs.3, 4 A number of drugs targeting PDF have been developed; however, some bacterial strains (including mutated resistant strains) exhibit different levels of response to these inhibitors in biological inhibition assays.5, 6 Therefore, additional design strategies for the development of stronger and more specific PDF inhibitors are warranted.

Recent architectural studies of PDF provides a framework for understanding the mechanism by which the peptide interacts with the interior of the ribosomal tunnel.7, 8 Nonetheless, many questions remain to fully understand the mechanism by which a new protein is processed and targeted, as well as the co-and posttranslational mechanisms required for the peptide to attain its final folded state.

In this study, we report the detailed crystal structure of the Enterococcus faecium PDF (EfPDF) complex with malonic acid and demonstrate the architectural basis for binding of the N-formyl polypeptide and for access of inhibitors to the active site of the enzyme. These structural studies will contribute to an improved understanding of the basis of peptide recognition and, thus, for antibacterial drug design.


Cloning, expression, and purification

EfPDF was amplified from a vancomycin resistant Enterococcus faecium (VRE) cDNA library using PCR with the following primers: 5′-GGG CAT ATG ATG ATT ACA ATG GAT GAT AT-3′ (forward, NdeI site underlined) and 5′-GGG CTC GAG CTA CTC GAT CAC CAA TAC GC-3′ (reverse, XhoI site underlined). The PCR products were ligated into the pET28a vector (Novagen). The supernatant of a E. coli cell lysate was loaded onto a HisTrap column (GE healthcare) to purify the enzyme. The protein was further purified by separation on a HiLoad 26/60 Superdex-200 prep-grade column (GE healthcare) using an elution buffer of 10 mM Tris-HCl (pH 8.0), 50 mM NaCl, and 2 mM DTT.

Crystallization, data collection, and structure determination

Crystallization of EfPDF was performed using the sitting-drop vapor-diffusion method at 22°C in 2.0 M Na-malonic acid, pH 4.0. X-ray diffraction data were collected from the cooled crystal in 30% (v/v) ethylene glycol as a cryo-protectant using beamline 6 C at the Pohang Light Source (PLS, South Korea). The raw data were processed and scaled using DENZO and SCALEPACK from the HKL2000 program.9 An initial model was obtained using molecular replacement. The MOLREP program10 in the CCP4 program suite was used with a model of E. pneumoniae PDF (PDB ID 1ML6). The structure was refined using the CNS program.11 Editing and adjustment of the model were carried out using a sigma A weighted 2Fo-Fc, Fo-Fc with the Coot program.12 Figures were generated using PyMOL13 and the final models were validated with PROCHECK.14 Statistics for data processing and refinement are listed in Table I.

Table I. Data Collection and Refinement Statistics
ParametersPDF-malonic acid complex
Data collection statistics 
 Space groupP6422
 Unit cell parameters (Å, °)a = b = 149.470, c = 143.329 α = β = 90°, γ = 120.0°
 Resolution range (Å)20 – 2.7 (2.8–2.7)
 Completeness97.3 (97.7)
 Redundancy7.8 (6.5)
 Average I/σ(I )12.22 (3.45)
 Rmerge (%)0.121 (0.304)
Refinement statistics 
 Rwork/Rfree (%)20.0/22.9
 r.m.s.d. bonds (Å)0.006
 r.m.s.d. angles (°)1.2
Ramachandran plot (%) 
 Most favored94
 Additionally allowed5
 Generously allowed1

Protein Data Bank accession number

The coordinate and structure factors for the E. faecium PDF have been deposited in the RCSB Protein Data Bank with the accession code 3CMD.


Crystal packing of EfPDF

The EfPDF crystal belongs to the hexagonal space group P6422 with unit-cell dimensions of a = b = 149.470 Å, c = 143.329 Å, and α = β = 90°, γ = 120.0°, with two monomers occupying the asymmetric unit. Each monomer is comprised of 187 residues, including 15 residues from an N-terminal expression tag, with a corresponding crystal volume per protein mass (Vm) of 5.00 Å3 Da−1 and a solvent content of 75.4%, a value well above the normal range observed for most proteins [Fig. 1(a)].15 An interesting feature of the EfPDF structure is that it includes an ordered N-terminal expression tag region that has access to the active site pocket of the neighboring molecule [Fig. 1(b)]. The unfolded expression tag peptide of the molecule has ∼82% of its total surface area buried, but these regions are not in contact with any residues in the cavity of the neighboring molecule.

Figure 1.

Crystal structure of EfPDF. (a) Crystallographic packing of 75.4% solvent content. Twelve of the EfPDF monomers generated a large 80 Å pore. (b) Unfolded N-terminal expression tag peptides of all molecules accessed the active site pockets of neighboring molecules. The expression peptide has ∼82% buried surfaced area, but makes no contacts. A 2Fo-Fc electron density map contoured at 1σ is shown in gray in the N-terminal region of the molecule. Molecules A and B are represented by green and cyan, respectively. The expression tag region is represented by dark-blue sticks. The surface of the active site pocket is represented in red. (c) Ribbon diagram of the EfPDF monomer structure. The unfolded expression tag peptide in the N-terminal region lies along the disordered region located between β2 and β3 (CD loop). Helices are represented in red, β strands in yellow, loops in green and the expression tag peptide loop in light blue. The active site metal ion is represented by the orange sphere. (d) Malonic acid interacts with the metal ion and Glu158, which is the catalytic residue. The active site pocket is loosely associated with the malonic acid. The active site residues are shown by ball-and-stick representations. The metal ion is represented by the orange sphere.

Overall structure and comparison with other PDFs

EfPDF is a classic α + β fold protein containing three α-helices, four 310-helices, an 8-stranded β sheet, and several long loops [Fig. 1(c)]. The molecule is disordered along the eight amino acids of the CD loop composed of Pro80–Pro87 between strands β2 and β3. This region is suggested to be the CD-loop “lid” that must be opened in order for substrate molecules to access the catalytic site and is therefore conformationally flexible.16, 17 The metal ion located in the active pocket of the enzyme is coordinated by His157 and His161 from the HEXXH motif of α3 and by Cys114 from the EGCLS motif in the long loop between the β5 and β6 strands.

In most PDF structures, the metal ion is coordinated by water and an absolutely conserved cysteine and two histidines, but this region of the EfPDF structure includes the hydroxyl group of the malonic acid [Fig. 1(d)]. The metal ion is pseudo-tetrahedrally coordinated by the Sγ atom of Cys114, the Nϵ2 atoms of His157 and His161, and the hydroxyl group of malonic acid. The four coordinated atom interactions are within ∼2.9, ∼2.5, ∼2.5, and ∼3.1 Å, respectively, of the metal ion. The hydroxyl atom of malonic acid interacts with the Oδ of Glu158 within ∼2.6 Å. Composite annealed omit maps were used to confirm the malonic acid molecule. The temperature factors of the malonic acid molecule (average = 40.2 Å2) are different from those of the main-chain atoms of the binding pocket (average = 20.6 Å2), indicating that it is loosely held in the pocket. Moreover, the non-interacting hydroxyl group of malonic acid has a different conformation and orientation [Fig. 1(d)]. The malonic acid is not an efficient inhibitor of PDFs, possibly as a result of dissociation due to the temperatures used for the structural analysis. In addition, we attempted to conduct binding studies by isothermal titration calorimetry (ITC); however, the results indicated that PDF has either no interaction or a very weak interaction with malonic acid (data not shown).

Nevertheless, the structure of the EfPDF-malonic acid co-crystal provides insight into inhibitor design. Previous PDF structural analysis results have shown that two water molecules in the active site pocket and a glutamic acid in motif II play a critical role in the deformylation catalysis mechanism.3 The EfPDF-malonic acid structure has similar atomic positioning, with the hydroxyl group of malonic acid located in the canonical water position and pseudo-tetrahedrally coordinated with metal ion at ∼3.1 Å [Fig. 1(d)]. In addition, the other hydroxyl group of malonic acid is positioned to interact with the Oϵ2 atom of Glu158 on the inner side of the S1′ pocket with a bond length of ∼2.5 Å [Fig. 1(d)]. Therefore, the hydroxyl groups of malonic acid are perfectly situated to prevent catalysis by the nucleophilic water and glutamic acid. We regard this atomic positioning of malonic acid to be critically informative for antibacterial drug design.

Suggested recognition of N-terminal formylated peptides on the basis of the crystal structure

In this report, we suggest that PDF specifically recognizes N-formyl polypeptides by using the unfolded N-terminal expression peptide to access the active site pockets of neighboring molecules [Fig. 1(b)]. PDF has a horizontal axis that forms a shallow “U” shaped cavity of approximately 15 × 10 Å, which is accommodated by the α3 helix and CD loop region in the EfPDF structure [Fig. 2]. Between each edge of the “U” shaped surface there is a distance of ∼11–18 Å. This data indicates that folded structures which have α-helices or β-sheets would be unable to effectively approach the active site pocket because α-helices and β-strands generally have diameters of 15 and 30 Å, respectively. If a folded structure approached, steric hindrance would occur and the highly mobile CD loop (disordered region) could interfere, whereas an unstructured polypeptide would be able to easily access the “U” shaped pocket. Access from below the “U” shape would be possible for a polypeptide, however access near the CD loop region may be prevented. Entrance through the side of the “U” shape is also possible, but would provide limited access to the active site pocket due to a strict edge and the high mobility of the CD loop region. Based on our structure, N-formyl polypeptides have three possible access routes: A, B, and C in Figure 2. In our structure, access route A is through a half tunnel of ∼13 Å. Access route B is through a half tunnel of ∼14 Å, but in a region located near the high mobility CD loop. Access route C is through a half tunnel of ∼19 Å and access of the N-formyl polypeptide to the active site pocket would be potentially unrestricted. To our knowledge, we have demonstrated here for the first time that the “U” shape and the dynamic nature of the CD loop prevent folded structures from accessing the active site of EfPDF. These results indicate that PDF recognizes unfolded N-formyl polypeptides, thus increasing substrate selectivity by preventing access by structured peptides.

Figure 2.

Proposed architectural mechanism of peptide recognition. Approximately 15 × 10 Å “U” shaped cavity in the EfPDF crystal structure. This shape prevents access by folded structures containing α-helices or β-sheets. Unfolded peptides can access the site from the side. The black dotted circle shows the region which prevents structurally folded proteins from entering, thus increasing the selectivity for unfolded peptides. The CD loop is located in the area of the red dotted circle. “X” indicates a route of entry that is likely to be inaccessible. Green ribbon is unfolded N-terminal expression tag peptides in our result.

Recently, novel antibiotics targeting PDF from vancomycin-resistant E. faecium have been screened.18 We believe that our crystal structural provides useful information for the design of strain-specific PDF inhibitors, based on the atomic coordination positions of malonic acid. Moreover, we suggest a mechanism for the specific recognition of N-formyl polypeptides by PDF and propose that this structural insight provides a framework for understanding the architectural mechanism of nascent protein processing.


The authors thank Dr. H. S. Lee, K. J. Kim, and K. H. Kim for assistance during data collection on beamline 6C of Pohang Light Source, Korea.